Twitter (X) has emerged as a critical platform for politicians to engage with voters, shape perceptions, and construct narratives, prompting extensive research into their Twitter behavior to analyze political discourse and sentiment, especially around elections. Studies have highlighted how opposition parties utilize Twitter for adversarial rhetoric against governing parties, but less explored is how shifts in government status affect the sentiment in politicians' tweets. This paper investigates whether transitions from opposition to government (or vice versa) influence Twitter sentiment, using Colombia’s 2022 electoral shift—where the left-leaning Pacto Histórico moved from opposition to power, and the center-right Centro Democrático transitioned to the opposition—as a case study, employing automated sentiment analysis and change point detection to examine changes in tweet sentiment following this political upheaval.

Tweets were scraped with the GetOldTweets-Python project. Sentiment analysis was performed with the pysentimiento Python library which contains a model trained on 5K Spanish tweets and change point detection methods were used to assess shifts in sentiment over time.